Q. How can the payroll data be negative simultaenously with a reduction in the unemployment rate?

A. Because they are distinctly different surveys. The payroll survey polls a set of employers, while the unemployment rate is derived from a survey of households.

That is why the payroll survey has shown 63,000 jobs lost, while the unemployment rate has dropped from 4.9% to 4.8%. A bear would say the unemployment rate is less credible (due to a higher statistical margin of error), while a bull would assume the payroll data to be unrepresentative, as it does not reflect small business (which is where most people work). In recent years, payroll numbers have had to be revised up. Way up, actually.

You are free to research further and decide. James Surowiecki has recently written a great primer.